Informed Consent and Decision Making Among Participants in Novel-Design Phase I Oncology Trials

Abstract

Background:

Although phase I clinical trials are the gateway to progress in cancer therapies, this setting poses ethical challenges to ensure that patients provide consent free from misunderstandings of therapeutic intent or unrealistic expectations of benefit. The design of phase I oncology trials has evolved rapidly over time and today includes more targeted agents and combinations of experimental drugs with standard drugs, which may further complicate how patients understand phase I research participation.

Methods:

We conducted semistructured interviews regarding motivations, decision making, and understanding of trial purpose nested within a phase I clinical trial of a novel PI3kinase inhibitor combined with a standard oral chemotherapy in 18 participants.

Results:

Fewer than half of patients correctly identified the safety and dosing objectives. The inclusion of a targeted agent was attractive to participants and was perceived as an indicator of less toxic or more efficacious therapy, with less appreciation for added risks. The significance of a cellular drug target, without a known predictive biomarker of response, was unclear to patients. The inclusion of a standard drug in the regimen attracted patients with more treatment options than traditional first-in-human participants. Patients frequently expressed a realistic understanding of prognosis and uncertainty of benefit, but simultaneous hopes for extraordinary outcomes.

Conclusion:

Novel phase I oncology trial designs may attract patients with less constrained treatment options, but the inclusion of targeted drugs and combinations including standard chemotherapies is likely to complicate understanding of safety and dosing objectives and likelihood of personal benefit for purposes of informed consent.

BACKGROUND

Inadequate participation in clinical trials is a well-recognized problem in oncology.^1-5 Early studies of safety and dosing, otherwise known as phase I trials, are the gateway to development of new therapies. When considering how to maximize recruitment to these trials, one key concern is the ethical balance between advancing new therapies and ensuring that phase I patients, who generally have advanced cancer and may have limited treatment options, provide informed consent.

Available studies suggest that phase I participants with cancer are often motivated by the hope of personal medical benefit^6-9 and that their expectations of this benefit are unrealistically high.^8,10-13 Participants report that they are adequately informed,^7,8,14 but studies testing trial-related knowledge have had conflicting results.^6,7,10 Communication with phase I candidates may be complicated by their strong desire to find effective therapy; one study found that none of the hypothetical adverse events proposed to patients, including a 10% risk of treatment-related death, significantly deterred the desire to participate,¹¹ and another found that over a quarter of patients allowed to choose their own dosing level chose the highest available unstudied dose.¹⁵ Patients often believe a phase I intervention will be effective despite specific statements by their physicians to the contrary.¹⁶ Physician difficulties with communication about phase I studies are also well documented, including presumptive decision making, use of highly complex language,¹⁶ and failure to discuss disease prognosis or treatment alternatives, including palliative care.^17,18

Hopes of phase I patients for personal medical benefit have raised ethical concerns of a therapeutic misconception.¹⁹ In the most concerning documented cases, the patient does not understand the fundamental differences between research and treatment; in phase I trials, patients may believe that they have a high likelihood of personal benefit from research designed to test safety and dosing rather than efficacy. Attempts to address such misconceptions through changes to the informed consent process have had limited success.^6,15,20

The current literature regarding decision making among phase I oncology trial participants has several key limitations. It largely describes populations facing traditional first-in-human trials using single-agent experimental therapies. These studies do not reflect the rapid evolution and increasing complexity of today’s phase I oncology trials, which include targeted therapies and oral drugs, as well as experimental/standard drug combinations, most of which are not first-in-human studies.²¹ It is unclear how historical ethical concerns regarding informed consent and therapeutic misconception should be applied in this environment. To address this research gap, we conducted semistructured interviews with participants entering a phase I study of an experimental PI3kinase inhibitor in combination with the Food and Drug Administration–approved chemotherapy capecitabine to identify key motivators and barriers to participation and to explore how they gathered, understood, and made use of information regarding trial purpose, risks, and benefits.

METHODS

Interviews were conducted with successive enrollees in an investigator-initiated phase I clinical trial at two academic medical centers. The trial tested the safety and dosing of a novel PI3kinase inhibitor, BKM120, in combination with capecitabine for metastatic breast cancer. Eligible participants were female, had received up to three prior chemotherapies for metastatic breast cancer, had normal organ function, and had not received prior capecitabine. Both drugs were administered orally in the outpatient setting; four cohorts received escalating doses of the experimental agent BKM120, and capecitabine was given at a therapeutic dose and schedule (2,000 to 2,500 mg/m²/d, 14 days on and 7 days off) to all participants. The study, including the interview portion, was approved by the University of North Carolina Institutional Review Board, and written consent was obtained. Eighteen of 25 study participants completed the interview within 7 days of treatment initiation. To minimize recall bias, respondents were interviewed on the first day of treatment if possible; two interviews were completed between days 2 and 7. Interviews were conducted by a single interviewer trained in qualitative techniques (K.E.R.-H.), using a 10-item interview guide (Fig 1). To ensure confidentiality and minimize undue influence, the interviewer was not involved in the participants’ clinical care. For two of the 18 respondents, telephone interviews were substituted for respondent convenience. The length of telephone and in-person interviews was similar. All interviews were audio-recorded, transcribed, and de-identified by redacting place and person names. Analysis was performed using Atlas.ti (version 7.1; Scientific Software Development, Berlin, Germany). Two coders (K.E.R.-H. and M.C.R.) independently coded all transcripts using a grounded theory approach. Before analysis, a preliminary code set was developed based on prior literature regarding factors in clinical trial participant decision making, the structure of the interview guide, and a conceptual framework on the basis of Becker’s Health Belief Model (Appendix Fig A1, online only).²² The code set was permitted to evolve in an inductive fashion by adding codes if additional material was identified that did not fit within the initial framework, and all transcripts were reviewed again for application of the additional codes. Discordance between coders was resolved by consensus. The final set comprised 35 codes. The codes with definitions, frequency of occurrence, and exemplary quotes are listed in Appendix Table A1 (online only). After coding, codes were grouped into five themes on the basis of the relatedness between pairs of codes in a correlation matrix as well as discussion between the coders (Fig 2). Because patients were asked a specific standardized question regarding the purpose of the trial, responses are reported as a proportion; other results are described qualitatively.

Fig 1.

Interview guide for study participants.

Fig 2.

Analytic codes grouped by theme.

RESULTS

Trial Purpose

Fewer than half of participants (seven of 18) specifically described the purpose of the trial as dose finding or safety, although the consent form stated that the purpose was to find the maximum safe and tolerable dose for the combination of the standard and experimental drugs. Most respondents said that the purpose of the study was to evaluate whether the combination was better than the standard treatment alone. Several participants explicitly acknowledged that finding an effective therapy was their own agenda but not the purpose of the study; one said, “the trial purpose is not my purpose.”

Motivators for Trial Participation

Most respondents identified the recommendation of their primary oncologist as a reason they participated. Most of these stated that they trusted their doctor’s judgment about what was best for them. Five respondents were encouraged to participate by a family member with medical training or work in a scientific field. Despite the availability of online resources, no patient identified online information seeking as a key part of their decision-making process, although one participant did use the ClinicalTrials.gov Web site. Rather, face-to-face communication with the primary oncologist, consulting academic physician, and/or research nurse was almost always described as the primary source of information regarding the trial, with written information as a secondary resource.

Some respondents were motivated to participate by their positive impression of the study team on their initial visit. One stated, “So, we came up here and talked to everybody. I got so much more information in that first day. It was just astounding. And I realized that I need more information, I need my questions answered.” This attention received from study personnel was perceived by some as a mark of good care.

Specific features of the trial design were key motivators for participation. The oral route of administration was highlighted as a benefit by most respondents, and only one patient saw this as a disadvantage. Several respondents identified the study combination of a novel therapy with capecitabine as the best of both worlds; they were able to receive a therapy with known efficacy plus a possibly more effective bonus drug. One patient described this as follows: “the effect that we were doing one drug that is considered standard of care and adding to it was really a huge factor for me because at this point, with my particular cancer, I didn’t want to do something that was completely experimental.”

Other motivators included the desire to help others with cancer, although this was not mentioned as a primary motivation by any respondent. Some interviewees described having dependent children as a specific motivator. Several described their cancer treatment as a fight, and two identified the trial as a way of being aggressive with the cancer. Other patients sought research participation because they perceived it as cutting-edge care or a novel alternative that was outside the box.

Alternatives to Participation

No patient described herself as having no other treatment options. Almost all patients described another standard chemotherapy agent as an option had they not joined the study. Interestingly, several participants specifically stated that they would have been treated with capecitabine off-study. One patient mentioned nutritional and complementary therapies as an alternative. No patient listed palliative care, hospice, or no additional cancer-directed treatment as an option.

Downsides of Participation

For respondents referred from other locations, several described travel to the academic center as a downside of trial participation. This concern manifested both as time away from family and as regret about leaving their familiar primary oncologist or treatment team. A few respondents mentioned a general dislike for the idea of research or feeling like a guinea pig. However, this trade-off was considered acceptable in exchange for access to better treatment; one participant said, “People think of research as you know, like, the guinea pig syndrome. You know, you can have these horrible side effects and stuff and for me, I was just willing to—I am willing to take that chance for the benefit.”

Perceptions of Risk and Benefit

When asked to recall adverse effects from the consent process, respondents acknowledged that adverse effects had been discussed with them and accurately listed a variety of adverse effects of both drugs, including hand-foot syndrome, cytopenias, depression, hyperglycemia, and greater toxicity of the standard drug when the experimental drug was added. However, respondents generally listed only one to three adverse effects and did not specify which drug was associated with which adverse effect. Minimizing language or humor was frequently used to deflect the adverse effect question. One respondent said she was told about a long list of adverse effects, and another said, “Well, everything has side effects.” Another joked, “Death. That’s the only one that really bothers me. The others I can handle.”

Most respondents expressed clinically reasonable hopes for the outcome of therapy, such as stopping progression or shrinking the tumor. Only two respondents stated that they hoped for cure as a possible outcome. Several respondents stated that the outcome was out of their control, either by referring to a specific religious faith or expressing a generally fatalistic viewpoint. One said, “I don’t worry about that because there’s absolutely nothing I can do about what’s going to happen but to work on what’s available now.” Some respondents specifically acknowledged that the treatment might not work as they hoped, but that hope was an important part of their approach to coping with advanced cancer. One said, “I’m not walking around with rose-colored glasses. But, you know, I’m also very hopeful.” Another explained that “my expectations and my hopes are two different things.”

Several respondents stated that their understanding was that metastatic breast cancer could not be cured, but then expressed hopes for benefit that essentially amounted to a cure. The word remission was used by some respondents to describe their hoped-for outcome; others said they hoped the cancer would go away. Some respondents specifically identified hope for long-term survival through future research advances. One stated, “No evidence of disease. Yeah. That’s what I’m hoping for. I understand, you know, when you become stage 4 there is no longer, cure is not an option. But remission and NED [no evidence of disease] is what you want to see happen. And that’s my goal…to stay alive while they continue to find out more research on breast cancer so that they can use better things.”

DISCUSSION

In this qualitative study of participants entering a phase I oncology trial testing the safety and dosing of a standard and experimental drug combination, we identified several key elements that may help to ensure adequate informed consent and optimal decision support.

In contrast to some previous studies of first-in-human phase I research, respondents in our study identified a number of perceived treatment options. Many selected the study because it represented an acceptable life fit rather than because of a lack of other alternatives, and only a few described themselves as having limited options. Although patients with more perceived treatment choices may be less subject to ethical concerns of coercion, they may also require assistance during the consent process in understanding how phase I study participation differs from other treatment options.

Understanding of the phase I trial purpose remained problematic in this study compared with prior literature, with fewer than half of participants correctly identifying dose-finding and safety end points. It is possible the inclusion of a standard chemotherapy drug (in this case, capecitabine) may complicate patient understanding of the short-term aims of phase I studies. In previous randomized trials, capecitabine for refractory metastatic breast cancer has shown objective response rates of 13% to 36% and progression-free survival of 3 to 6 months,^23-27 and is an accepted therapeutic option. Patients intuitively and correctly perceived that they had a reasonable likelihood of benefit from the standard portion of the regimen and logically concluded that if an experimental agent was added, the researchers’ goal must be to ascertain whether the combination was better than the prior standard. Some respondents thought of the perceived long-term research goal (ultimately finding, through a series of investigations, whether the new combination is better than the standard therapy) as the purpose of the phase I trial, rather than the immediate aims of determining dosage and safety. However, even among respondents who correctly identified the short-term trial objectives, almost all identified personal goals for participation that were distinct from the trial’s goals, and many stated that they were participating to obtain clinical benefit. This duality has been described by Agrawal and Emanuel as understanding versus appreciation²⁸; patients may rationally understand that the trial has a limited chance of benefit or a dose-finding purpose, yet decide to participate for personal reasons.

This study also identified confusion regarding the targeted nature of BKM120. Some respondents perceived increased likelihood that the drug would work against their cancer, although no biomarker predictive of response is currently available for PI3kinase inhibitors. For other respondents, the description of a drug target reassured them that this treatment was likely to be less toxic than standard therapies because it selectively targeted cancer cells; in reality, the experimental compound does have potential toxicities distinct from those of capecitabine, including hyperglycemia and depression, which were described in the consent documents and which would not be eliminated by successful targeting of the PI3kinase receptor. To our knowledge, this is the first study to report on how phase I participants incorporate the concept of targeted therapy into their assessments of treatment risks and benefits.

We found that almost all respondents had a technical understanding of their incurable prognosis. However, several respondents expressed hopes for outcomes similar to cure, such as remission, no evidence of disease, or surviving until a curative therapy could be found. These conversations may be examples of optimism bias, in which patients have expectations for their own outcome that substantially exceed those they expect for the average patient with their condition. Jansen et al²⁹ have argued that optimism bias may compromise the voluntariness of informed consent. However, several respondents explicitly acknowledged being realistically prepared for poor outcomes, but also hoping for exceptional outcomes, which some saw as an important aspect of fighting cancer. Miller and Joffe^30,31 have argued that such optimism represents a coping strategy that may be adaptive and does not invalidate informed consent. A recent quantitative study suggested that higher expectations of trial benefit are associated with dispositional optimism, or overall positive life outlook, rather than optimism bias, a finding that supports Miller’s proposition.³²

Our study is exploratory and has several key limitations. The respondents were uniformly female and had breast cancer, which has multiple treatment options in the metastatic setting. These experiences may not reflect the concerns of patients with other tumor types. Several of the issues discussed here may not be salient to other phase I trial designs, such as first-in-human or single-agent trials, or those that enroll a population with a specific predictive marker of treatment response.

In conclusion, novel phase I clinical trial designs are attractive to patients and may offer participation opportunities for patients with more treatment options and less vulnerability to coercion than traditional first-in-human populations. However, the inclusion of standard anticancer drugs in phase I trials may make it more complicated for patients to understand that phase I studies are designed to test safety and dosing rather than efficacy. Furthermore, phase I trial designs containing drugs described as targeted may require additional education regarding the persistent potential for toxicity, and the limited information regarding benefit from hitting the target, during informed consent. Our findings confirm earlier research showing that patients with advanced cancer may hold both realistic expectations for prognosis and simultaneous hopes for extraordinary outcomes, and that they value hope as part of their cancer journey. Additional research should focus on the development of more effective tools to help oncology providers communicate, and patients understand, the goals and complexities of modern phase I oncology trials, while respecting the patient’s autonomy and desire to maintain hope during the last phases of life.

Appendix

Fig A1.

A conceptual model for phase I patient decision-making process on the basis of Becker’s Health Belief Model

Table A1.

Codebook With Definitions, Examples, and Frequencies

AUTHOR CONTRIBUTIONS

Conception and design: Katherine E. Reeder-Hayes, Elizabeth C. Dees

Collection and assembly of data: Katherine E. Reeder-Hayes

Data analysis and interpretation: All authors

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Informed Consent and Decision Making Among Participants in Novel-Design Phase I Oncology Trials

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/journal/jop/site/misc/ifc.xhtml.

Katherine E. Reeder-Hayes

No relationship to disclose

Megan C. Roberts

No relationship to disclose

Gail E. Henderson

No relationship to disclose

Elizabeth C. Dees

Consulting or Advisory Role: Novartis (I)

Research Funding: Novartis, Genentech/Roche, Bayer, Pfizer, Merck, Eli Lilly, Cerulean Pharma